Signal to Noise Ratio

[Jamal]

Hi all,

I am trying to calculate signal to noise ratio of a photonic analog-to-digital converter using an electrical carrier analyser.
According to the formula SNR = 6.02N+1.76 , it shouldn’t go beyond 61.96 dB but it is giving an SNR of 103dB at output.
Can someone point out the mistake please and is it the correct method of calculating SNR ?

Attachments:

1. 2GHz-multiwavelength-multiadc-no-filter.osd

[Damian Marek]

In your simulation there are zero sources of noise. The noise at the carrier analyzer is then -100 dBm giving you the very large SNR of 103dB. I understand that the ADC introduces a distortion to the signal, but the way it is set up right now it does not introduce any noise and therefore the SNR is unbounded.

To calculate this SNR from the distortion you will have to inspect the RF Spectrum and compare the power of the in band signal to the out of band signal, which I am guessing is > 50 GHz.

[Jamal]

Thanks Damian for clearing that out
Can you please explain a little more, like how do I inspect both the bands separately from the spectrum ?

[Damian Marek]

Visually inspecting the RF Spectrum Analyzer you already have attached there will do the job. You can use a marker to find the peak power of the signal at 0 GHz and the approximate noise power at 50 GHz and compare the SNR.

The ADC component only allows for the addition of thermal noise. The only other distortion it brings into the signal is the lower sampling rate.

[Jamal]

Also, is the Analog to Digital converter component in optisystem 14 an ideal ADC?
And is there any way to model clock jitters or non-linearity issues in ADC ?

[Jamal]

Thank you Damian,
But I am confused as the SNR of ADC as described in the application report by texas instruments describes SNR as ratio between the power of signal to the average power of the noise floor which is due to the quantization noise and thermal noise of ADC.
http://www.ti.com/lit/an/slaa510/slaa510.pdf

By looking at the spectrum I can find the peak of individual frequencies but how to calculate average for a noise floor ?

Also you wrote ‘ the way it is set up right now it does not introduce any noise and therefore the SNR is unbounded ‘ so I have added noise also but it does not show up at the output.

Attachments:

1. 2GHz-multiwavelength-multiadc-no-filter-noise.osd

[Jamal]

Can anyone tell the full scale range of this ADC ?

[Ranjeet Kumar]

Hi,
I think the above formula for calculating SNR is only for PCM modulation formats. Since optisystem 14 used ideal component by default but we can change its parameters to see variations.

[Jamal]

I have added noise in the components, but the noise doesn’t show up in the spectrum.
Can anyone suggest the problem ?

Attachments:

1. 2GHz-multiwavelength-multiadc-single-input-10bit1.osd

[Author]

Posts